Federated learning communication method and device

Abstract

The embodiment of the invention discloses a federated learning communication method and device, and the method comprises the following steps: a plurality of modeling participation devices sending a modeling request to a central server, wherein the modeling request carries target information; receiving grouping information that the central server divides the plurality of modeling participation devices into a plurality of participation groups according to the target information; determining proxy communication equipment of the participation group by utilizing an algorithm in the participation group; sending the encryption parameters to the proxy communication devices of the participation groups, so that the proxy communication devices of the plurality of participation groups send the encryption parameters to a central server; and receiving a processing result of the encryption parameter from the central server sent by the proxy communication device of the participation group. Therefore, according to the embodiment of the invention, the central server only performs network communication with the proxy communication equipment of each participation group, so that the overall network communication overhead is saved, the network communication high load of the central server is relieved, and the network communication efficiency is improved.

Description

technical field [0001] The present application relates to the field of computers, in particular to a federated learning communication method and device. Background technique [0002] Federated Learning (Federated Learning) is an emerging basic technology of artificial intelligence, which is used to solve the problem of updating models of terminal equipment. Under the premise of standardization, efficient machine learning can be carried out among multi-participants (devices) or multi-computing nodes. Generally speaking, a federated learning communication network includes multiple devices: a central server and multiple devices participating in federated learning, where the central server can be called a coordinator, and multiple devices can be called multiple participants. In the process of building a model through federated learning, multiple devices need to continuously establish communication connections with the central server to communicate, that is, the central server f...

AI Technical Summary

Problems solved by technology

[0004] In order to solve the problem of establishing a model of federated learning in the prior art, the network communication load of the central server is very high, and the communication overhead of the entire network communication is very large, resulting in network communication In order to solve the problem of low efficiency, this application provides a federated learning communication method, which can save the overall network communication overhead and alleviate the high network communication load of the central server

Images